回归分析与灰色系统耦合用于水环境预测研究

Prediction of water environment based on coupling technique of pluralistic regression and grey system

探讨了将多元回归分析模型和灰色系统（ＧＭ（１，１））耦合应用于武汉市东湖水体污染预测的方法和技术。首先根据多年污染因子浓度监测资料分别建立了ＣＯＤ、ＢＯＤ、ＴＮ、ＴＰ与人口、经济和捕鱼量之间的多元线性回归方程；之后根据人口、经济和鱼产量的历史资料，用ＧＭ（１，１）模型拟合，再反推得到预测值；最后，将人口、经济和鱼产量的预测值代入多元线性回归方程得到污染物浓度的未来值。这样将环境污染与社会经济联系起来更能反映环境质量的变化趋势。结果表明，到２１ 世纪，东湖水体的ＴＮ、ＴＰ、ＣＯＤ、ＢＯＤ 的浓度将分别是１９８２ 年的２．５８、３．７０、２．４５ 和７．８９ 倍；比１９９２年分别增长３３．５％ 、６８．１８％ 、６８．２８％ 和２２９．２４％ ，将处于超富营养状态，如不及时治理，生态环境将进一步恶化。

The coupling technique of pluralistic regression and Grey System (GM<1,1>) was discussed and applied to water pollution prediction of the East Lake in Wuhan city in China. Firstly, pluralistic regressions were established respectively between COD, BOD, TN, TP and population, economy and fish yield; Secondly, the history material of population, economy and fish yield was linked and calculated with GM<1,1>,so the prediction value of population, economy and fish yield was obtained; Finally the prediction value of population, economy and fish yield was input to the pluralistic regression of COD, BOD, TN and TP, and the prediction value of the contaminant content was get The prediction results of the East Lake were satisfactory In this way, environment pollution was combined with social economy, the prediction could reflect objectively the tendency of the water contaminant change. If the East Lake was not cleaned in time, in the 21 th century, the densities of TN, TP, COD, BOD in it would be 2 58, 3 70, 2 54 and 7 89 times as those in 1982 and increase by 33 50%, 68 18%, 68 28% and 229 24% when compared with that in 1990. If so, organic pollution and utrophication will be more and more serous day by day